← Back to Skills Marketplace
654
Downloads
0
Stars
1
Active Installs
6
Versions
Install in OpenClaw
/install zongjie
Description
将问题解决过程整理成结构化文档,沉淀经验并保存到指定记忆文件,便于后续查阅和复盘。
Usage Guidance
技能本身是方法论/文档化脚本,行为与描述一致。但在安装或启用前请注意:
- 核实 ~/.openclaw/scripts/memory_mgr.py 等本地脚本的内容,确认它们不会意外上传数据到外部端点;
- 注意要保存的日志或配置可能包含敏感信息(API token、凭据、个人数据),在写入记忆前清理或脱敏;
- 确认本地 openclaw CLI 和脚本的来源可信,必要时限制文件权限或在隔离环境中运行;
- 允许代理自主调用技能是平台默认行为:如果你不希望技能自动运行,请通过平台设置限制其调用权限。
Capability Analysis
Type: OpenClaw Skill
Name: zongjie
Version: 1.0.5
The `SKILL.md` file explicitly instructs the AI agent to execute shell commands, specifically `python3 ~/.openclaw/scripts/memory_mgr.py add <agent> "<内容>" P1 L1` and similar commands in the '常用命令' section. These commands pass user-derived content (e.g., `"<内容>"`) as arguments directly to a shell. This design introduces a significant shell injection vulnerability, allowing a malicious user to potentially execute arbitrary commands on the host system if the input is not properly sanitized or quoted. While this presents a critical RCE risk, there is no clear evidence of intentional malicious behavior (e.g., data exfiltration, backdoor installation) by the skill developer, thus classifying it as suspicious rather than malicious.
Capability Assessment
Purpose & Capability
技能名称与描述是“整理并保存问题解决过程到记忆文件”。SKILL.md 中的所有路径(如 memory/topics/...、~/.openclaw/scripts/memory_mgr.py、~/.openclaw/openclaw.json)和示例命令(openclaw CLI、python3 调用 memory_mgr.py)都与该目的直接相关,所需操作(读取/合并现有记录、写入记忆文件、更新索引)是合理且一致的。
Instruction Scope
说明文档限定了要保存的场景、模板和步骤,未授权地扩展行为的指令较少,整体范围与目标一致。但 SKILL.md 示范性地引用了读取本地配置和日志(例如 ~/.openclaw/openclaw.json、openclaw logs、memory_mgr.py),这些操作会访问本地文件与日志,可能包含敏感信息。建议在允许技能写入/读取前,确认要访问的具体文件和内容范围。
Install Mechanism
这是纯说明型技能(无 install 规范、无代码文件要下载),因此不会在安装阶段写入磁盘或下载第三方二进制,安装风险很低。
Credentials
技能不要求任何环境变量或外部凭据,这是与其用途相称的。需要注意的是,技能建议操作和示例会访问本地 OpenClaw 配置和日志文件(~/.openclaw/...),这些文件可能包含令牌或凭据;访问这些文件对技能目的有意义,但应在可信环境下进行。
Persistence & Privilege
技能没有设置 always:true,也不请求修改其他技能或系统范围的配置。它建议使用本地脚本保存记忆(memory_mgr.py),这些都是局部、可审计的操作,权限要求与其描述相符。
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install zongjie - After installation, invoke the skill by name or use
/zongjie - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.5
- 优先级与生命周期调整:P0 改为“永不过期”,P1 由 30天延长为 90天,P2 由 7天延长为 30天。
- 优先级说明更明确,P0 说明添加“关键偏好”。
- 未更改模板或功能用例,仅优化和细化文档描述。
v1.0.4
- 增加了基于 OpenViking 的三层记忆管理系统和优先级(P/L)说明
- 模板更新:支持在总结中标记优先级与层级(例:`> P1 | L1`)
- 新增记忆分类、层级和生命周期定义表
- 增补常用命令说明,便于记忆添加和查询
- 拓展保存规则,支持更多类型归档(如新功能)
- 强化输出内容,保存时告知分类与存储层级
v1.0.3
- Enhanced the summary template to include "关键信息" (key info like logs, commands, links) and expanded solution steps.
- Updated instructions to extract and record troubleshooting details in more depth.
- Expanded example with detailed commands, logs, and key configuration paths.
- Clarified that solution steps should be written out explicitly and key debug artifacts should be captured in the summary.
v1.0.2
- Introduced step M3: 合并同类项,实现与已有记录的合并与更新。
- 将后续步骤顺序调整,M4: 保存分类,M5: 输出确认。
- 更新示例,展示如何合并新问题到已有记录。
- 其他内容未变。
v1.0.1
- 精简事件总结模板,更易书写和保存关键信息
- 明确仅保存有参考价值的重要技术/经验事件,排除问答、查询、闲聊内容
- 优化步骤和分类,简化整理和归档流程
- 统一输出,告知用户保存位置即可
- 新增案例,便于理解操作流程
v1.0.0
zongjie 1.0.0 changelog:
- Initial release: Document and structure event postmortem workflows
- Provides clear steps for extracting key info, formatting reports, and saving to memory folders
- Includes template for troubleshooting documentation with timeline, root cause, solutions, results, and summary
- Details guidance on when and how to use the skill for maximum future reference
Metadata
Frequently Asked Questions
What is 总结?
将问题解决过程整理成结构化文档,沉淀经验并保存到指定记忆文件,便于后续查阅和复盘。 It is an AI Agent Skill for Claude Code / OpenClaw, with 654 downloads so far.
How do I install 总结?
Run "/install zongjie" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is 总结 free?
Yes, 总结 is completely free (open-source). You can download, install and use it at no cost.
Which platforms does 总结 support?
总结 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created 总结?
It is built and maintained by marklyk (@marklyk); the current version is v1.0.5.
More Skills